Analysis Of Healthy Breeding Behavior Of Dairy Cows Based On Dynamic Perception Data Mining

With the level of urbanization and the continuous improvement of people’s living standards,people’s demand for dairy products has gradually increased.In order to meet the growing demand for dairy products and improve the economic benefits of dairy farmers,timely and accurate estrus testing is necessary.The traditional small-scale farms mainly adopt artificial breeding methods.The growth status and breeding behavior of dairy cows depend on the naked eye observation and manual recording of the staff,which requires a lot of time and manpower,especially for scale intensive farming.The field has a huge task and a high false positive rate.With the wide application of Internet of Things technology in animal husbandry,high-definition cameras,pedometers,infrared body temperature detectors and other equipment are gradually being used in modern pasture farming.The wide application of Internet of Things equipment in the large-scale breeding process of dairy cows has accumulated a large number of individual data and video surveillance data of dairy cows.Through the effective use of these dairy cow behavior indicator system data,the behavioral characteristics of dairy cows that affect the healthy breeding of animals can be captured in time,and an early warning model of dairy cows’ healthy reproductive behavior will be established.Dairy farming can be run in an orderly manner.The economic benefits of farmers will increase.And The modern dairy farming industry can develop healthily.This paper analyzes the milk production data and exercise data of dairy cows.Because cows are excited and violent during estrus and the milk yield is reduced,the data of cows in estrus will be different from normal.The DBSCAN clustering algorithm can not only find clusters of arbitrary shapes,but also can be used for detection of outliers.The key to the implementation of the algorithm is to determine the appropriate parameters.This paper considers the influence of different index variables on the estrus behavior of dairy cows,and uses the entropy method to determine the weight values of different indicators.The appropriate neighborhood radius is determined based on the weight using the weighted distance,and the value of MinPts is determined by the stability of the clustering result.In the final clustering result,the points that are assigned to the same cluster class are detected as non-estrus points.Points that are not assigned to the cluster class are detected as estrus points.In order to verify the effect of clustering,this paper uses variance analysis to analyze the estrus data and non-estrus data.For the data of milk production,P=0.001<α=0.05,the difference between the mean value of the estrus sample data and the non-estrus sample data is significant.That is,whether estrus has a significant effect on the milk production observation.For the data of exercise amount,P=0<α= 0.05,the difference between the mean value of the estrus sample data and the non-estrus sample data is significant.That is,whether estrus has a significant influence on the exercise amount observation value.The above results illustrate the feasibility and effectiveness of the study.